MCP server

Connect Claude, Codex, Cursor, and other MCP-compatible AI agents to Quickly’s hosted Streamable HTTP MCP server. The server is included on every plan, including the free tiers.

Connection

Use a workspace API key to connect an MCP client to Quickly. The server uses the same permission model as the public API.

Server URL: https://mcp.askquickly.ai/api/mcp/v1
Authentication: Authorization: Bearer qk_...

Open MCP setup Create an API key View API reference MCP server pricing

Connect from Codex

Create a workspace API key in Quickly, store it as an environment variable, then register the Streamable HTTP MCP server.

export QUICKLY_API_KEY="qk_..."

# ~/.codex/config.toml
[mcp_servers.quickly]
url = "https://mcp.askquickly.ai/api/mcp/v1"
bearer_token_env_var = "QUICKLY_API_KEY"

Tools

The v1 MCP catalog supports reading knowledge, inspecting records, triggering playbooks, and auditing outcomes. Destructive actions require the right scope and still pass through Quickly’s workflow safety controls.

Tool	Description
`knowledge_search`	Search workspace knowledge bases and return grounded results.
`knowledge_base_list`	List available workspace knowledge bases.
`playbook_list`	Find playbooks an agent can run.
`playbook_trigger`	Trigger a governed playbook run with approval and audit controls.
`playbook_run_get`	Fetch a specific playbook run.
`playbook_run_list`	List recent playbook runs.
`entity_list`	Search workspace entities and CRM-style records.
`entity_get`	Fetch a specific entity record.
`entity_create`	Create a workspace-scoped entity record.
`entity_update`	Update a workspace-scoped entity record.
`task_list`	List recent Quickly task and audit records.
`task_get`	Fetch one task with execution details.
`agent_list`	List scheduled playbook agents.

API key scopes

Assign only the scopes the connected agent needs. Available MCP key scopes are:

knowledge:read
knowledge:write
playbooks:read
playbooks:execute
logs:read
agents:read
messages:read
messages:write
web:read

Agent-readable discovery

AI agents can start with these public discovery files to find the API, MCP server, setup instructions, and safety model without crawling the full marketing site.

/llms.txt
Curated AI-readable site map
/llms-full.txt
Expanded API and MCP context
/openapi.json
Machine-readable REST API surface